Method and apparatus for recurrent auto-encoding
a technology of auto-encoding and recurrent encoding, applied in the field of signal encoding, can solve the problem of reducing the audio quality of the corresponding decoded signal
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Benefits of technology
Problems solved by technology
Method used
Image
Examples
Embodiment Construction
[0019]Devices and methods to perform signal encoding are described. In a particular example, a recurrent autoencoder architecture, referred to as a feedback recurrent autoencoder (“FRAE”), is described with application to the compression of sequential data with temporal correlation. The recurrent structure of FRAE is configured to efficiently extract the redundancy embedded along the time-dimension of sequential data and enable compact discrete representation of the data at the bottleneck in a sequential fashion.
[0020]In a particular example, an audio signal corresponds to a temporally correlated signal x={x1, x2, . . . , xt, . . . }, for which xt is referred to as signal x in time frame t. An autoencoder includes an encoder portion and a decoder portion. The encoder portion includes a first neural network and a quantizer. The decoder portion includes a second neural network. A simple architecture of an autoencoder follows a forward-only design that treats each data point xt indepen...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


